The prior art has not heretofore recognized that the form of temperature cycle employed for high temperature processing can have any effect on the rate of oxygen precipitation.
In the early days of semiconductor manufacturing, the form of temperature cycle employed for high temperature processing was not known to have any significance with regard to the semiconductor wafer. To treat a wafer at a given temperature, the wafer was merely thrust into a hot furnace for the requisite time and then pulled out.
However, it was soon realized that this form of temperature cycle was responsible for cracks or warpage in the semiconductor wafer. Subsequent developments have attempted to control wafer warping or cracking by modifying the form of the temperature cycle to slowly heat the semiconductor ("ramp up") prior to high temperature processing and slowly cool the semiconductor ("ramp down") after high temperature processing. For example, in U.S. Pat. No. 3,723,053 to Myers et al. entitled "Heat Treating Process for Semiconductor Fabrication", the yield of devices from a semiconductor wafer is increased by determining the critical temperature gradient that causes stress above the yield point of the semiconductor material and controlling both the rate of heating and cooling so that the temperature gradient is lower than the critical temperature gradient. Similar forms of ramp up/ramp down temperature cycles have been widely adopted for controlling wafer cracking or warping. However, no relation has been recognized between the ramp up/ramp down form of temperature cycle and the rate of oxygen precipitation.